rec.autos.simulators

(clip)

At least the figure for perception of frame rate has risen from 24 to
30 :)  People have claimed on here that you can't see a difference in
anything over 24 fps!  If you've never done any experiments, or read
an article, how do you pull a figure of 30 fps from the air?

I think there is a point where higher frame rates won't make driving
any easier, probably at around the 30 fps mark.  That would have more
to do with human reflex time than with perception of the eye.  If you
are getting information faster than your ability to react to it, then
it should be OK.  From my experience, though, 50 fps would definitely
look better than 30 fps, and I can see a difference between the two.

The most important thing would be to _never_ drop below 30 fps.

--
Pat Dotson
IMPACT Motorsports
http://www.racesimcentral.net/

I don't think u can see the different beyond 30fps. Your eyes can't tell the
diffenences unless u have cat eyes ;-)

Yes you can (see my earlier post for details).

- Eric

Says who? At 30fps my friend and I are equal in Quake. If I have 60fps
and he has 30fps I kick his ***big time, and if he has 60fps he does
the same to me. I'm not too worried if it can be scientifically proven
that it makes a difference, the fact that the results improve is good
enough for me as it is results I need, not proof.

Then again, If you don't find any difference it's all the more money
you can save, so lucky you! :-)

Cheers!
John

hehe, yeah but that's not saying much <GGG>

Mike.

Are you sure that's because you see better at 60 fps? Or is it because the
program can accept your inputs faster at 60 fps? I'm almost one hundred % sure
the latter is the case. When a computer draws 30 fps, the time required for the
input to be processed is two times the amount of time when drawing 60 fps. The
response time is doubled from 60 fps to 30 fps. A big response time makes you
worse at Quake playing, NOT the amount of frames per second...

When that is said, I'd like to point out that the response time is also very
important in racing sims, and that a low response time is to be strived for. My
point is that the frame rate is in a way an indicator of how pressed the cpu is,
not an indicator of how well the picture is perceived by the eye. The eye hardly
notices the difference between 30 and 60 fps, but the "feel", the response does.

There are other ways to decrease the response time, for instance using a graphics
board that demands less of the cpu (I think one of the best chips must be the
Rendition V2100 in this respect). A low frame rate can be symptoms of
1) A hard pressed cpu, that can't keep up the triangle feed to the graphics board

2) A hard pressed graphics board
Of these cases, the no.1 creates the worst response, simply because the cpu also
processes the input...

I find a difference, but there is much more to this than if a game has 60 fps or
not! As I said earlier; By some strange reason, GPL gives around 30 - 35 fps and
physics sim frequency 288 hz and NOT 120 fps and 20 hz physics calculation... In
my view, Papyrus is spot on, spending the cpu cycles on physics rather than
fps... Besides, GPL prooves that fill rate isn't everything to a racing sim.

---Asgeir---

Your PS/2 and serial ports only sample the mouse at about 40 and 44Hz
respectively, and Quake2 only samples the controller input at 10Hz.

- Eric

I believe it's higher than that...read on.

The other thing that isn't being discussed is that if you can generate
50 new frames per second (with independent vehicle positions for
each), the images that you'll see will be 1/50th of a second
"outdated" or extrapolated to their actual displayed time.

If you only see 30 per second, you'll have more of a delay, or require
more extrapolation in the displayed data, neither of which contributes
positively to car control... :-)

---Jim

What possible good would it do to display more frames than you had
done game engine calculations for (as some of the frames would
necessarily be pixel-perfect duplicates assuming a deterministic
renderer)?

In your second example, I'd argue that the TRUE framerate could only
be 50Hz.

---Jim

Not in Papyrus sims. Papy sims sample the input devices and calculate
physics at a rate independent of the graphics frame rate. So, in the
case of Papy sims, inputs don't get accepted any "faster". What
happens is you can have a tighter feedback loop between the output
(the generated frames) and the physics engine, allowing you to drive
more precisely.

Does it really matter the precise physiological effects, or is the
fact that the feel of the game improves sufficient? I personally
believe you're dead wrong on the physiological effects, but I also
believe that's largely unimportant, since the real issue is whether
the game play is qualitatively better at higher frame rates. I believe
it is.

---Jim

On Wed, 23 Sep 1998 12:40:09 +0100, "Mike Laskey"

It was YOU I was talking about wide-boy, and I thought I was being
generous.... ;-)

Cheers!
John

Bingo - that's it precisely.

As I described before about Quake, what I seek is not proof of
frame-rate improving my visual feedbakc or whatever, the fact that
there is a difference in socre is sufficient. What causes that
difference is arguable for whoever is interested, but while they're
doing that I've probably got all the goodies and too busy enjoying
myself :-)

Cheers!
John

Yes, of course, I should have said "than 75 fps and 75hz physics sim",
that would make more sense...

Of course, just another case of a little brain fade...

---Asgeir---

Yes, that must be the way to express it.

I agree, I really do. My only point was that there is more to a sim than
the framerate, although badly put (but I guess we can all agree on that
anyway, so I didn't have to argue on that).

But, when it comes to the physiological effects, there seems to be
extremely little hard facts, and I'd like to know if anyone has ever read
or seen anything documenting eye perception, because I do NOT accept
sayings like: "I play better quake in 75 fps than 30 fps, therefore my eye
can tell the difference between 75 and 30 fps...". But, as you rightfully
say, the feel in the racing sim counts no end, if this feel is created in
the eye or in the hand, that's completely unimportant as long as the feel
is good.

---Asgeir--- /Totally out of bounds on knowledge, but who isn't?

The input sampling makes control less precise at higher framerates in
Quake, thus the popularity of m_filter to mask this as well as cl_maxfps
to cap the max framerate at 35-40 which is below the sampling rate for a
normal PS/2 and serial mouse (with a USB mouse this can be overcome due
to it's 120Hz sampling rate).  Those that play better at 75fps do so as
a result of the visual framerate increase, in spite of the control
deficiency.

- Eric